mays

Plugin mays from marketplace dotfiles in stevenmays/dotfiles. Commands read from the repo's manifest — 5 days ago.

Skills (16)

• adversarial-review

  Two-family adversarial review of a plan or a diff — one fresh Claude subagent and one external run (GPT through Codex, or Grok through Cursor) on identical briefs, launched concurrently, then triaged into verified blockers, concerns, and rejected findings. Use before approving a plan, before shippin

• claims-verification

  Check the factual claims a plan makes about existing code against the repo — extract atomic claims, have a fresh subagent answer neutral questions without seeing the plan, then score each claim. Use after a plan is drafted and before adversarial review, or whenever a workflow asks to verify a plan's

• claude-subagents

  Dispatch planning, review, implementation, or a scoped question to a fresh Claude subagent with the Agent tool — which subagent type and model to pick, what the prompt must carry, and how to check what comes back. Use whenever a workflow hands work to a Claude subagent, including the Claude half of

• codex-image-generator

  Default image generation path. Generate or edit raster images (hero images, illustrations, textures, product shots, mockups, transparent cutouts) through the Codex CLI's built-in image_gen tool, run in a subagent. Use whenever the user asks for an image, picture, illustration, or visual asset and do

• deslop

  Clean up AI slop in existing code — measure complexity with real tools, delete dead and redundant code, refactor the worst hotspots first, and prove behavior is unchanged. Use when the user asks to clean up slop, deslop, reduce complexity, untangle spaghetti, simplify a god function, or make AI-gene

• external-subagents

  Run a planning, research, review, or implementation task on a second model family as a background subagent that cannot hang the session — GPT through the Codex CLI, or the newest Grok through the Cursor CLI where Codex is missing. A wrapper kills stalled runs and salvages partial output. Use wheneve

• extreme-code-quality-review

  Run an extremely strict maintainability review for abstraction quality, giant files, and spaghetti-condition growth. Use for an extreme code quality review, deep code quality audit, or especially harsh maintainability review.

• frontend-craft

  Process for pushing AI-built UI past generic output — seed-string variety, an independent design-critic loop with screenshots, generated images and video instead of CSS-only visuals, and a subtraction pass that removes AI tells. Use with the frontend-design skill, not instead of it, whenever buildin

• gemini-image-generator

  Generate or edit raster images with Google's Nano Banana model through the Antigravity CLI's built-in generate_image tool, run in a subagent. Use when the user names Gemini, Nano Banana, or Antigravity. Not for SVG, icons, or code diagrams. The default image path is codex-image-generator.

• gemini-subagents

  Run a planning, research, review, or implementation task on a Gemini model as a background subagent through the Antigravity CLI (`agy`), behind a wrapper that kills stalled runs and salvages partial output. Use when a workflow names Gemini, Argon, Antigravity, or `agy` as the second model. Not for g

• model-check

  Model-check the failure paths of a design with Quint before it ships — races, crashes, redelivery, duplicate delivery, out-of-order events, and timeouts. A fresh subagent writes the spec and its code mapping, mutants prove each invariant can fail, and the caller runs every check. After implementatio

• serverless-aws

  Patterns for AWS Lambda, DynamoDB, SQS, and Secrets Manager. Use when working on serverless AWS projects.

• skill-optimizer

  Use when optimizing, benchmarking, or improving an existing skill that works but is inconsistent — runs an autonomous mutation-and-scoring loop with binary evals, keeps improvements, discards regressions, and produces a changelog. Triggers include "optimize this skill", "improve this skill", "benchm

• ste-writing

  Google developer documentation style fused with Simplified Technical English (ASD-STE100) discipline, adapted for software engineering, plus economy rules that curtail verbose drafts. Use whenever writing or rewriting technical prose for any reader. Covers documentation, tutorials, READMEs, runbooks

• teach-me

  Teach a technical concept to mastery. Teach it first on an interactive HTML lesson page, take questions, practice in chat, quiz with confidence ratings, drill gaps, re-quiz, and schedule spaced reviews. Use for learn, teach me, quiz me, review.

• writing-style

  Write in Steven's voice—pragmatic, curious, pedagogical. Opens on what actually happened, builds mental models from first principles, uses worked examples, and handles uncertainty honestly. Use only when the user asks for an essay, blog post, or article. A tutorial, README, or explainer belongs to s

Slash commands (10)

• /mays:babysit-pr

  Keep your open PR merge-ready — fix CI failures, address review comments, resolve merge conflicts

• /mays:diagnose-bug

  Root-cause a bug methodically — build a red-capable repro loop first, then bisect, hypothesize, fix, and lock it down with a regression test

• /mays:extreme-code-quality-review

  Extremely strict maintainability audit of the current branch (code-judo, 1k-line rule, spaghetti growth)

• /mays:fix-merge-conflict

  Resolve all merge conflicts non-interactively, then build and test

• /mays:onboard-repo

  Day-one bootstrap for a new codebase — CLAUDE.md and a permissions allowlist in one run

• /mays:pre-review

  Review the current branch for correctness, AI slop, and repo-convention drift before opening a PR

• /mays:review-pr

  Review someone else's PR — findings as STE-written Conventional Comments, with optional posting to GitHub

• /mays:security-audit

  Audit home-directory dotfiles for security issues — package cooldown config, secrets, permissions, shell hygiene, supply-chain settings

• /mays:ship

  Run the full pre-PR golden path — convention review, correctness review, tests — then commit, push, and open the PR

• /mays:test-and-fix

  Run tests and fix failures until green

Subagents (1)

• extreme-code-quality-review

  Extreme code quality audit (maintainability, structure, 1k-line rule, spaghetti, code-judo). Dispatched via the Agent tool after a parent gathers the diff and changed-file contents. Reads its rubric from the `extreme-code-quality-review` skill's `SKILL.md`.

Hooks (3)

• PreToolUse

  ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/block-destructive-git.sh

• PreToolUse

  ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/block-hardcoded-secrets.sh

• PostToolUse

  ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/post-edit-format.sh

Similar plugins